Hacker Newsnew | past | comments | ask | show | jobs | submit | nodesocket's commentslogin

Is there just the one Mowin' Man in More After Dark?


Optimus ins’t gonna be a master welder, electrician, HVAC technician, general contractor anytime soon.


Debatable. Optimus will be programmed by watching thousands of YouTube videos and by indexing the entire NEC. It is entirely possible that a single Optimus can be a master welder, electrician, and HVAC technician. What Optimus can't do is take responsibility for the completion of the work as being done to code. Nor is it clear how much Optimus will cost to buy or operate.

Underestimating humanoid capabilities is a mistake; the real challenge will be the economics and the flow of responsibility. But imagine a future where every structure built has a pov video documenting its construction. I don't like the implication for the trades any more than I like the implication of Claude for putting me out of work either.


Latest LLMs that have good visual understanding (Gemini and GPT) are able to tell if construction work is up to code. You can give it a picture of electrical wiring, plumbing, framing and it will tell you if it's up to code with very good accuracy.


Ok, but if you think that's true, then it would be a huge mistake to put Optimus on two legs instead of a boom lift.


If anybody is looking for alternatives, I’m a very happy user of Bezel (https://beszel.dev) on my home lab servers including a Raspberry Pi Kubernetes cluster and TrueNAS. Highly recommend.


Same, built on pocketbase too. And rather mature with strong docs now.


I also can feel a speed/sluggishness difference between my M1 Max MacBook Pro and my M4 Pro (14 core).


Honest question: Which one feels faster?


Obviously the M4 Pro feels faster.


It’s not obvious to me. It’s somewhat surprising, even, given the specs I just looked up:

M1 Max: 400 GB/s memory bandwidth, ~20.8 FP16 Tflops

M4 Pro: 273 GB/s memory bandwidth ~17.2 FP16 Tflops


True but memory bandwidth is not the entire story. Geekbench (while not a end-all benchmark) shows:

                      Single-Core      Multi-Core
  M1 Max, 10-core     2,204            14,004
  M4 Pro, 14-core     3,311            24,711

Single-threaded CPU: M4 Pro is roughly 1.5× as fast. Multi-threaded CPU: M4 Pro is roughly 1.75× as fast. GPU: roughly equal to a full 32-core M1 Max, with the M4 Pro slightly ahead in Geekbench Metal.


Makes sense. I am still getting a feel for the different levels of these processors.


Upgrading to 48GB memory is $600…. Gut punch. I have a M4 Pro (top cpu skew 14 cores). Any idea the performance gain to M5 Pro? I feel like they holding off releasing M6 Pro because a new studio is dropping soon as well.


At time of his talk Bitcoin was trading at around $8. Let that sink in.


There a few factors at play. I believe China has launched a very successful campaign of misinformation in order to slow down US progress and allow them to “win” the “AI War”. Then climate activists and let’s be frank far left leaning people have taken the bait hook line and sinker and amplified it.

Second Anthropic at the start was very vocal about datacenter doom. Their motivate besides wanting regulatory capture is again slow down their competitors and corner the market.

Lastly mainstream media loves a good doomer story of tech people getting rich displacing working class Americans while “destroying earth”.


Paying Typora user here, and wondering the same.


This sort of coddling and weak policy is setting students up for a rude awaking when they get into the professional workforce.


Companies don't expect interns or new graduates to perform at the level of a senior engineer. They expect and plan that they have to ease them into the "professional workforce" somewhat.

Anyways, the biggest tech companies in Michigan are automakers, Dow chemicals, and Whirlpool. These are all respectable places and have different expectations than X or OpenAI or companies like that.


Seeing the purpose of education to be preparing you to be part of "the professional workforce" is the problem.


Seems useful for s3 buckets. I suppose a script that watches for new objects, calculates its sha256 and stores that in a DB, then checks for duplicate hashes would be a fairly trivial task. Though s3 doesn’t support symbolic links so accounting would need to be handled by server side code.


While the idea of data-deduplication does apply everywhere, this specifically depends on the APFS CoW data feature, so it's not applicable outside of macOS (or APFS Volumes)


The same concepts exist in ZFS, BtrFS and, IIRC, XFS.


also, worst case you can do this with link features? Which is basically every system under the sun.


but link is dangerous, as a edit made to one file will then be made to all copies.

For example, If I have two files /home/m/todo.txt and /home/m/backup2025/todo.txt that are identical and you then replace /home/m/todo.txt with a link to /home/m/backup2025/todo.txt and I later edit /home/m/todo.txt then backup copy would also be updated. With apfs clones the backup would not be modified when I updated the file.


Link might be desirable for your e-book or mp3 collections. Or .iso images, MAME ROMs, and other files that are unlikely to be changed.


Links are not CoW


I don't think it's very useful but there's already a checksum in the meta data of an S3 object: https://docs.aws.amazon.com/AmazonS3/latest/API/API_Object.h...

It might be better to keep an index of paths to checksums and use the checksum as object key in S3.

I'm not sure what you like to achieve but if you use restic with S3 as a backend you can achieve much better deduplicate and compression.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: